Kamis, 25 Mei 2017

Accuracy of ICU Predictions Often Varies

Accuracy of ICU Predictions Often Varies


The ability of intensive care unit (ICU) clinicians to predict accurately the mortality and functional outcomes of patients 6 months after discharge is variable and reflects, in part, the confidence those clinicians have in their predictions, the authors of a new study write.

Prognostic accuracy was greatest when physicians and nurses agreed on their predictions and had high confidence in them, Michael E. Detsky, MD, from the Palliative and Advanced Illness Research Center, University of Pennsylvania Perelman School of Medicine, Philadelphia, and the Sinai Health System and the University of Toronto, both in Toronto, Ontario, Canada, and colleagues write in an article published online May 21 in JAMA.

This finding may influence planning for postdischarge care and clinicians’ discussions with patients and/or their families and caregivers, they explain.

It also underscores the uncertainty that often exists when making prognoses for ICU patients.

The data “inform and affirm the approach to presenting prognosis used by many ICU clinicians,” Jesse B. Hall, MD, from the Section of Pulmonary and Critical Care Medicine, University of Chicago, Illinois, writes in an accompanying editorial. Dr Hall advises these clinicians to acknowledge to themselves and the families they counsel that ICU patients often are in “a prognostic gray zone.”

The researchers conducted a prospective cohort study of patients in five ICUs across three hospitals in the University of Pennsylvania Health System. Eligible patients were enrolled in their third through sixth day of ICU care if they had received mechanical ventilation for more than 48 consecutive hours, vasoactive infusions for more than 24 consecutive hours, or both.

The participating clinicians included the patient’s attending physician and primary bedside registered nurse. They were asked to predict the patient’s chances of in-hospital survival, plus five more outcomes at 6 months: survival, return to original residence, independent toileting, ability to independently climb a flight of stairs, and cognitive function. For each prediction, the clinicians also were asked to rate their confidence in the prediction, on a Likert scale of 1 (not confident at all) to 5 (very confident). All predictions were made within 24 hours of patient enrollment in the study.

A total of 47 physicians and 128 nurses participated in the study, making predictions for 303 patients, who had a median age of 62 years (interquartile range, 53 – 71 years) and included 173 (57%) men. Six-month outcomes were verified for 299 of those patients.

For each predicted outcome, the researchers calculated likelihood ratios (LRs) as well as several other operating characteristics. Higher positive LRs suggest greater accuracy at predicting adverse outcomes, whereas lower negative LRs reflect greater accuracy at predicting favorable outcomes.

Physicians were most accurate at predicting 6-month mortality (positive LR, 5.91 [95% confidence interval (CI), 3.74 – 9.32]; negative LR, 0.41 [95% CI, 0.33 – 0.52]), and problems with independent toileting (positive LR, 6.00; 95% CI, 3.18 – 11.30). They were least accurate at predicting cognitive outcomes (positive LR, 2.36 [95% CI, 1.36 – 4.12]; negative LR, 0.75 [95% CI, 0.61 – 0.92]).

Nurses most accurately predicted in-hospital mortality (positive LR, 4.71; 95% CI, 2.94 – 7.56), independent toileting (negative LR, 0.48; 95% CI, 0.30 – 0.78), and ability to climb a flight of 10 stairs (negative LR, 0.48; 95% CI, 0.31 – 0.74). Similar to the physicians, they were least accurate at predicting cognitive function (positive LR, 1.50 [95% CI, 0.86 – 2.60]; negative LR, 0.88 [95% CI, 0.73 – 1.06]).

Physicians rated 41% to 55% of their predictions as 4 or 5 on the Likert scale, indicating high confidence. Nurses gave similar ratings for 44% to 57% of their predictions.

When physicians felt confident, “their predictions of 6-month survival, return to original residence, toileting independently, and ambulating up 10 stairs were significantly more accurate than when they were not confident,” the authors write.

Among nurses, “confident predictions of 6-month survival and return to original residence were significantly more accurate than their nonconfident predictions of these outcomes.”

Physicians and nurses both agreed and felt highly confident for 22% to 33% of predictions. In those instances, discriminative accuracy was “typically excellent, especially for 6-month mortality (positive LR, 40.35 [95% CI, 5.73-284.28], negative LR, 0.18 [95% CI, 0.06-0.50]) and toileting (positive LR, 15.75 [95% CI, 4.04-58.94]; negative LR, 0.11 [95% CI, 0.02-0.68]).”

Adding the clinicians’ predictions to prediction models such as the Acute Physiology and Chronic Health Evaluation III significantly improved the predictive accuracy of those models, the researchers add.

Clinicians should acknowledge that the accuracy of their predictions was “typically modest,” even when it was greater than chance, the authors warn. Also, given that accuracy was greatest when physicians and nurses agreed and felt confident, “ICU physicians should consider both their own confidence and their agreement, or lack thereof, with bedside nurses in deciding how to frame their prognostic judgments to patients and surrogates.”

In addition, the author recommend clinicians focus on 6-month mortality and functional outcomes in their discussions with families, as those outcomes seem to be associated with the most accurate predictions.

These findings show that “concordance of prognosis among clinicians — even at a single point and early in the course — is associated with the greatest accuracy,” Dr Hall writes.

However, he warns, the artificial conditions of a study such as this one do not reflect how clinicians make real-life decisions. Usually, they consider a patient’s unique set of circumstances plus their own knowledge and experience to assess the risk level for that particular patient, which they then communicate to the patient’s family.

Understanding how clinicians calibrate their judgments is “of paramount importance,” he concludes, “because patients and families deserve the most realistic assessment of outcome, properly framed by the reality that ICU clinicians can never offer perfectly accurate predictions.”

One study author reported receiving a grant from the National Institute of Nursing Research. Dr Hall reported receiving textbook royalties from McGraw-Hill and honoraria from the American College of Chest Physicians. The remaining authors have disclosed no relevant financial relationships.

JAMA. Published online May 21, 2017. Article full text, Editorial full text

For more news, join us on Facebook and Twitter



Source link

Tidak ada komentar:

Posting Komentar